CN109670467A - A kind of forest change method for quickly identifying based on SG filtering - Google Patents
A kind of forest change method for quickly identifying based on SG filtering Download PDFInfo
- Publication number
- CN109670467A CN109670467A CN201811595924.5A CN201811595924A CN109670467A CN 109670467 A CN109670467 A CN 109670467A CN 201811595924 A CN201811595924 A CN 201811595924A CN 109670467 A CN109670467 A CN 109670467A
- Authority
- CN
- China
- Prior art keywords
- forest
- filtering
- data
- ifz
- pixel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The present invention relates to a kind of forest change method for quickly identifying based on SG filtering, by the way that initial data is weighted the forest characteristics index IFZ that summation operation obtains describing the pixel Yu forest departure degree, it is recycled using the characteristic construction noise reduction of SG filtering, rebuild the time series array of IFZ within the scope of search time, according to the new time series and weight that continuously SG is filtered three times, judge whether to meet stopping criterion for iteration, if being unsatisfactory for, it is then made iteratively SG filtering, until meeting stopping criterion for iteration;At this time because SG filtering while filtering out noise may insure that the shape of signal, width are constant, and SG filtering be high frequency signal section, i.e. noise spot, smothing filtering can eliminate issuable interference.Solving the problems, such as that forest characteristics index IFZ noise is more causes errors of analytical results big.
Description
Technical field
The present invention relates to dynamic changes of forest resources monitoring technical fields, and in particular to a kind of forest change based on SG filtering
Change method for quickly identifying.
Background technique
It is periodically fixed that forest resource monitoring refers to that the quantity to the forest reserves, quality, spatial distribution and its utilization obstacle carry out
The work such as analysis, observation and the evaluation of position.By the research changed to forest litterfall, the quantity of the forest reserves can be understood in time
And quality, grasp the growth and decline changing rule and trend of the forest reserves, analyzing influence and nature, economy and the society for restricting Forest Growth
Meeting objective condition, establishs or updates files of forest resources, to the variation of announcement ecosystem environment and revegetation and rebuilds layout
Etc. being of great significance.With the rapid development of social economy, the rate of expansion of urban size gradually becomes faster, agricultural land area
It is continuously increased, causes area of woods smaller and smaller.Forest monitors the dynamic analysis for being conducive to forest, predicts that the variation of forest becomes
Gesture monitors and is formulated for area of woods forest conservation policy and provides foundation.
In terms of Monitoring on Dynamic Change algorithm, there are many scholars to propose for the remote sensing image of various different spatial resolutions
The change detection algorithms of many land cover patterns and vegetation.These variation detections generally divide the forest image of different times
Change information is extracted in analysis.Wherein forest characteristics index value IFZ is a kind of common forest Monitoring Index.But the forest characteristics
Index value IFZ has more noise, and it is larger to easily lead to errors of analytical results.
Summary of the invention
The purpose of the present invention is to provide a kind of forest change method for quickly identifying based on SG filtering, for solving forest
The more problem for causing errors of analytical results big of characteristic index IFZ noise.
In order to solve the above technical problems, the technical solution of the present invention is as follows:
The present invention provides a kind of forest change method for quickly identifying based on SG filtering, comprising the following steps:
Several forest images in a continuous time period are obtained, every width forest image is corresponding time point in monitoring range
Forest image;
For each pixel of each width forest image, forest characteristics index IFZ is calculated, to obtain each forest pixel
One group of original I FZ value in continuous time period;
To one group of original I FZ value of each forest pixel, first SG filtering is carried out, it will be in first filtered data
The data of each time point are compared to the data at corresponding time point in original I FZ value, take minimum value composition therein first
Reconstruct data;Simultaneously according to the size of data of first filtering front and back, the weight in each pixel each year is calculated;
Data will be reconstructed for the first time and carry out SG filtering again, by the data of each time point in filtered data again and original
The data at corresponding time point are compared in beginning IFZ value, and minimum value composition therein is taken to reconstruct data again;
SG filtering is carried out again to new reconstruct data, by the data of each time point in another filtered data
It is compared to the data at corresponding time point in original I FZ value, takes minimum value composition therein reconstruct data again;
According to the new time series and the weight that continuously SG is filtered three times, judge whether to meet iteration ends
Condition, if not satisfied, SG filtering is then made iteratively, until meeting the stopping criterion for iteration;
After meeting the stopping criterion for iteration, result is exported;It is indicated whether on judging each pixel at every point of time gloomy
Woods pixel, to judge forest change situation in monitoring range.
Beneficial effects of the present invention:
By the way that initial data is weighted the forest characteristics that summation operation obtains describing the pixel Yu forest departure degree
Index IFZ is recycled using the characteristic construction noise reduction of SG filtering, rebuilds the time series array of IFZ within the scope of search time, according to
The continuous new time series and weight that SG is filtered three times, judges whether to meet stopping criterion for iteration, if not satisfied, then
It is made iteratively SG filtering, until meeting stopping criterion for iteration;At this time because SG filtering may insure letter while filtering out noise
Number shape, width it is constant, and SG filtering be high frequency signal section, i.e. noise spot, smothing filtering can be eliminated can
The interference that can be generated.Solving the problems, such as that forest characteristics index IFZ noise is more causes errors of analytical results big.
Further, the weight of each pixel is expressed from the next:
Wherein, i indicates value of the pixel in 1 year, diFor IFZtrWith IFZ0Distance, dmaxFor diMaximum value;It calculates
Weight is to calculate iteration ends discriminant parameter.
Further, when judging whether to meet the stopping criterion for iteration, parameter F need to be distinguished according to iteration endskSentence
It is disconnected, FkIt indicates are as follows:
Wherein,For the IFZ value that the pixel reconstructs after kth+1 time filtering, FkIt is whole for the filtered iteration of kth time
Only distinguish parameter, m is research duration;It facilitates deciding on and whether meets stopping criterion for iteration, improve the accurate of operation in method
Degree.
Further, work as Fk≤Fk-1And Fk≤Fk+1When, iterative cycles are terminated, final result is exported;It improves in method and transports
The order of accuarcy of calculation.
Further, the mode for obtaining several forest images in a continuous time period is to pass through satellite or flight
Device is taken photo by plane;Be conducive to rapidly and accurately obtain forest image.
Detailed description of the invention
Fig. 1 is flow chart of the method for the present invention;
Fig. 2 be the embodiment of the present invention IFZ after SG is filtered effect diagram.
Specific embodiment
To keep the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, to the present invention
It is described in further detail.
Basic principle of the invention are as follows:
By the way that initial data is weighted the forest characteristics that summation operation obtains describing the pixel Yu forest departure degree
Index IFZ is recycled using the characteristic construction noise reduction of SG filtering, rebuilds the time series array of IFZ within the scope of search time, according to
The continuous new time series and weight that SG is filtered three times, judges whether to meet stopping criterion for iteration, if not satisfied, then
It is made iteratively SG filtering, until meeting stopping criterion for iteration.Wherein selecting SG filtering is because the maximum feature of SG filtering exists
It is constant in shape, the width that may insure signal while filtering out noise, signal width, that is, time Statistic analysis, this two o'clock
It is extremely important in the analysis of next step.And SG filtering be high frequency signal section (trend term), i.e. noise spot has
Conducive to reduction error.For the angle of forest change, forest change is slow process, and smoothing filter can eliminate possibility
The interference of generation.
Flow chart of the method for the present invention is as shown in Figure 1, specific method step are as follows:
1) forest auxiliary information is obtained;Taken photo by plane using satellite or aircraft obtain in a continuous time period several are gloomy
Woods image, every width forest image are the forest image at corresponding time point in monitoring range;It is gloomy by every using geographic assistant data
The image element extraction of woods image comes out;
2) standardization of data;The standardization of data uses following formula:
In formula, FZnIt is the index value after standardization;MnAnd SnIt is that wood land spectral value is averaged in wave band n respectively
Value and standard deviation, bnFor certain pixel the wave band spectral value.
3) forest characteristics index IFZ is calculated;For each pixel of each width forest image, forest characteristics index is calculated
IFZ, to obtain one group original I FZ value of each pixel in continuous time period;Forest characteristics are acquired according to the following formula to refer to
Number IFZ:
In formula, NB is used wave band sum, FZnFor corresponding standardization index value after wave band n processing.
4) first SG filtering, calculates weight;To one group of original I FZ value of each forest pixel, first SG filtering is carried out, it will
The data of each time point are compared to the data at corresponding time point in original I FZ value in first filtered data, take it
In the first reconstruct wavefront of minimum value composition after size of data, calculate the weight in each pixel each year;Weight is expressed from the next: number
According to;Simultaneously according to first filter
Wherein, i indicates value of the pixel in 1 year, diFor IFZtrWith IFZ0Distance, dmaxFor diMaximum value.
Data will be reconstructed for the first time carries out SG filtering again, it will each time point (unit: year) in filtered data again
Data be compared to the data at corresponding time point in original I FZ value, take minimum value therein composition to reconstruct data again;It is right
New reconstruct data carry out again SG filtering, by the data of each time point in another filtered data and original I FZ
The data at corresponding time point are compared in value, take minimum value composition therein reconstruct data again;According to continuously SG three times
Obtained new time series and the weight is filtered, judges whether to meet stopping criterion for iteration, is judging whether to meet iteration
When termination condition, parameter F need to be distinguished according to iteration endskJudgement, FkIt indicates are as follows:
Wherein,For the IFZ value that the pixel reconstructs after kth+1 time filtering, FkIt is whole for the filtered iteration of kth time
Only distinguish parameter, m is research duration.
Stopping criterion for iteration are as follows: work as Fk≤Fk-1And Fk≤Fk+1When, iterative cycles are terminated, final result is exported.If discontented
Foot is then made iteratively SG filtering, until meeting stopping criterion for iteration;For example, the new time sequence obtained after SG is filtered three times
Column are unsatisfactory for stopping criterion for iteration, at this time using the filtered data of third time SG as first reconstruct data, will reconstruct number for the first time
According to SG filtering (the 4th filtering) is carried out again, by the data of each time point in filtered data again and original I FZ value
In the data at corresponding time point be compared, take minimum value composition therein to reconstruct data again;It is another to new reconstruct data
Secondary progress SG filtering (the 5th filtering), will be in the data of each time point in another filtered data and original I FZ value
The data at corresponding time point are compared, and take minimum value composition therein reconstruct data again;According to continuously SG is filtered three times
The new time series and the weight that (third time, the 4th time, the 5th time) obtains, judge whether to meet stopping criterion for iteration,
If being unsatisfactory for condition, using the 5th filtered result as first reconstruct Data duplication above-mentioned steps.
After meeting stopping criterion for iteration, result is exported.
As shown in Fig. 2, the result exported after above-mentioned filtering for some pixel.Top broken line is virgin forest feature
Index IFZ, lower section broken line are forest characteristics index IFZ after SG iterative filtering, in figure there may be the numerical value of noise (such as
2009 and 2014), it will smoothly to eliminate interference, the variation tendency of script image is maintained, convenient for next step
Analysis.
5) judge that each pixel indicates whether forest pixel on (each year) at every point of time, to judge monitoring range
Interior forest change situation;The threshold value in coupling relationship tracing algorithm is set, it is special to the forest in step 4) in each point search time
Sign index IFZ noise reduction value is analyzed one by one.When this year the point value be greater than threshold value when, assert the point be non-forest, be less than threshold
Value then regards as forest, and the specific time of record forest change to non-forest analyzes the change type that may occur, obtains the picture
The relevant information of first forest change.In the variation time for counting each point, analyze the variation that the region forest may occur.For example, false
If forest characteristics index IFZ noise reduction value has exceeded given threshold at 2014 in Fig. 2, then recognizing the point is non-forest, is then analyzed
Forest change analyzes the change type that may occur to the specific time of non-forest.
This method is in research, to obtain more accurately as a result, it is desirable to each pixel point on image all in accordance with upper
The method of stating is handled.
This method obtains describing the gloomy of the pixel and forest departure degree by the way that initial data is weighted summation operation
Woods characteristic index IFZ is recycled using the characteristic construction noise reduction of SG filtering, rebuilds the time series number of IFZ within the scope of search time
Group judges whether to meet stopping criterion for iteration according to the new time series and weight that continuously SG is filtered three times, if discontented
Foot is then made iteratively SG filtering and is filtered noise reduction to forest characteristics index IFZ, until meeting stopping criterion for iteration;It solves
The forest characteristics index IFZ noise more problem for causing errors of analytical results big.
Claims (5)
1. a kind of forest change method for quickly identifying based on SG filtering, which comprises the following steps:
Several forest images in a continuous time period are obtained, every width forest image is the gloomy of corresponding time point in monitoring range
Woods image;
For each pixel of each width forest image, forest characteristics index IFZ is calculated, so that it is continuous to obtain each forest pixel
One group of original I FZ value in period;
To one group of original I FZ value of each forest pixel, first SG filtering is carried out, it will be each in first filtered data
The data at time point are compared to the data at corresponding time point in original I FZ value, take the first reconstruct of minimum value composition therein
Data;Simultaneously according to the size of data of first filtering front and back, the weight in each pixel each year is calculated;
Data will be reconstructed for the first time and carry out SG filtering again, by the data of each time point in filtered data again with it is original
The data at corresponding time point are compared in IFZ value, and minimum value composition therein is taken to reconstruct data again;
SG filtering is carried out again to new reconstruct data, by the data of each time point in another filtered data and original
The data at corresponding time point are compared in beginning IFZ value, take minimum value composition therein reconstruct data again;
According to the new time series and the weight that continuously SG is filtered three times, judge whether to meet stopping criterion for iteration,
If not satisfied, SG filtering is then made iteratively, until meeting the stopping criterion for iteration;
After meeting the stopping criterion for iteration, result is exported;Forest picture is indicated whether on judging each pixel at every point of time
Member, to judge forest change situation in monitoring range.
2. the forest change method for quickly identifying according to claim 1 based on SG filtering, which is characterized in that described each
The weight of pixel is expressed from the next:
Wherein, i indicates value of the pixel in 1 year, diFor IFZtrWith IFZ0Distance, dmaxFor diMaximum value.
3. the forest change method for quickly identifying according to claim 2 based on SG filtering, which is characterized in that be in judgement
It is not no when meeting the stopping criterion for iteration, parameter F need to be distinguished according to iteration endskJudgement, FkIt indicates are as follows:
Wherein,For the IFZ value that the pixel reconstructs after kth+1 time filtering, FkIt is distinguished for the filtered iteration ends of kth time
Other parameter, m are research duration.
4. the forest change method for quickly identifying according to claim 3 based on SG filtering, which is characterized in that work as Fk≤Fk-1
And Fk≤Fk+1When, iterative cycles are terminated, final result is exported.
5. the forest change method for quickly identifying according to claim 1-4 based on SG filtering, which is characterized in that
The mode for obtaining several forest images in a continuous time period is to be taken photo by plane by satellite or aircraft.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811595924.5A CN109670467A (en) | 2018-12-25 | 2018-12-25 | A kind of forest change method for quickly identifying based on SG filtering |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811595924.5A CN109670467A (en) | 2018-12-25 | 2018-12-25 | A kind of forest change method for quickly identifying based on SG filtering |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109670467A true CN109670467A (en) | 2019-04-23 |
Family
ID=66146088
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811595924.5A Pending CN109670467A (en) | 2018-12-25 | 2018-12-25 | A kind of forest change method for quickly identifying based on SG filtering |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109670467A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080165246A1 (en) * | 2007-01-06 | 2008-07-10 | Samsung Electronics Co., Ltd. | Method and apparatus for controlling intra-refreshing in a video telephony communication system |
CN104851087A (en) * | 2015-04-17 | 2015-08-19 | 华中农业大学 | Multi-scale forest dynamic change monitoring method |
CN106803059A (en) * | 2016-12-02 | 2017-06-06 | 浙江工业大学 | A kind of remote sensing vegetation index time series forest monitoring method |
CN106960119A (en) * | 2017-03-06 | 2017-07-18 | 南京信息工程大学 | A kind of fusion DI forest disturbance comprehensive characteristics index and method |
CN107886498A (en) * | 2017-10-13 | 2018-04-06 | 中国科学院上海技术物理研究所 | A kind of extraterrestrial target detecting and tracking method based on spaceborne image sequence |
CN108388866A (en) * | 2018-02-27 | 2018-08-10 | 海南热带海洋学院 | A kind of vegetation monotone variation trend-monitoring method and relevant apparatus |
-
2018
- 2018-12-25 CN CN201811595924.5A patent/CN109670467A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080165246A1 (en) * | 2007-01-06 | 2008-07-10 | Samsung Electronics Co., Ltd. | Method and apparatus for controlling intra-refreshing in a video telephony communication system |
CN104851087A (en) * | 2015-04-17 | 2015-08-19 | 华中农业大学 | Multi-scale forest dynamic change monitoring method |
CN106803059A (en) * | 2016-12-02 | 2017-06-06 | 浙江工业大学 | A kind of remote sensing vegetation index time series forest monitoring method |
CN106960119A (en) * | 2017-03-06 | 2017-07-18 | 南京信息工程大学 | A kind of fusion DI forest disturbance comprehensive characteristics index and method |
CN107886498A (en) * | 2017-10-13 | 2018-04-06 | 中国科学院上海技术物理研究所 | A kind of extraterrestrial target detecting and tracking method based on spaceborne image sequence |
CN108388866A (en) * | 2018-02-27 | 2018-08-10 | 海南热带海洋学院 | A kind of vegetation monotone variation trend-monitoring method and relevant apparatus |
Non-Patent Citations (8)
Title |
---|
JIN CHEN等: "A simple method for reconstructing a high-quality NDVI time-series data set based on the Savitzky–Golay filter", 《REMOTE SENSING OF ENVIRONMENT》 * |
LU SIQI等: "Spatio-temporal dynamics of shifting cultivation in Upland Myanmar using time series images and implications for REDD+", 《SPIE ASIA-PACIFIC REMOTE SENSING》 * |
吕莹莹: "基于 VCT 模型的森林干扰机制研究", 《中国优秀硕士学位论文全文数据库 农业科技辑》 * |
吴健生 等: "基于图像阈值法的森林雪灾损失遥感估测", 《地理科学进展》 * |
尹凌宇: "基于GF-1WFV数据的森林覆盖变化检测方法研究", 《中国优秀硕士学位论文全文数据库 农业科技辑》 * |
梁文海 等: "基于面向对象方法的 GF-2 影像桉树林信息提取", 《浙江农林大学学报》 * |
王颖洁 等: "基于时序Landsat数据的三江平原植被地表类型变化遥感探测研究", 《遥感技术与应用》 * |
黄春波 等: "基于时间序列统计特性的森林变化监测", 《遥感学报》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Li et al. | Adaptively constrained dynamic time warping for time series classification and clustering | |
CN110287932B (en) | Road blocking information extraction method based on deep learning image semantic segmentation | |
US10706260B2 (en) | Analyzing digital holographic microscopy data for hematology applications | |
Gong et al. | Change detection in synthetic aperture radar images based on image fusion and fuzzy clustering | |
US11587222B2 (en) | Method and apparatus for detecting defect pattern on wafer based on unsupervised learning | |
Kim et al. | Color–texture segmentation using unsupervised graph cuts | |
US9064151B2 (en) | Device and method for detecting plantation rows | |
US20100092075A1 (en) | Method of directed pattern enhancement for flexible recognition | |
CN110021425A (en) | A kind of relatively detector and its construction method and cervical cancer cell detection method | |
Zhang et al. | A new time series representation model and corresponding similarity measure for fast and accurate similarity detection | |
CN108664838A (en) | Based on the monitoring scene pedestrian detection method end to end for improving RPN depth networks | |
Yue et al. | Texture extraction for object-oriented classification of high spatial resolution remotely sensed images using a semivariogram | |
Ngugi et al. | A new approach to learning and recognizing leaf diseases from individual lesions using convolutional neural networks | |
CN116416884B (en) | Testing device and testing method for display module | |
CN110969121A (en) | High-resolution radar target recognition algorithm based on deep learning | |
Tu et al. | Hierarchical structure-based noisy labels detection for hyperspectral image classification | |
Yang et al. | Real-time classification for Φ-OTDR vibration events in the case of small sample size datasets | |
Liang et al. | Automatic defect detection of texture surface with an efficient texture removal network | |
CN103955711B (en) | A kind of mode identification method in imaging spectral target identification analysis | |
Bai et al. | Achieving better category separability for hyperspectral image classification: A spatial–spectral approach | |
Wang et al. | High-level background prior based salient object detection | |
Bionda et al. | Deep autoencoders for anomaly detection in textured images using CW-SSIM | |
Zeybek | Inlier point preservation in outlier points removed from the ALS point cloud | |
Li et al. | Hierarchical clustering driven by cognitive features | |
CN111488851A (en) | Traceability detection method, device, equipment and medium for fruit production place |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190423 |
|
RJ01 | Rejection of invention patent application after publication |